Browse Prior Art Database

Method to Generate Hypothesis From Social Networks and the Internet Disclosure Number: IPCOM000243845D
Publication Date: 2015-Oct-20
Document File: 3 page(s) / 861K

Publishing Venue

The Prior Art Database


Disclosed is a system to automatically search social media, identify different people diagnosed with a health condition, identify the date of diagnoses for each person, collect a set of characteristics for each person, and run machine learning in order to create a sorted list of correlated characteristic for the health conditions. The list generates a set of hypothesis that enables researchers to perform formal research and clinical studies.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 01 of 3

Method to Generate Hypothesis From Social Networks and the Internet

Some medical researchers have limited resources when researching the cause of disease. For example, a breast cancer researcher generally selects intuitive causes of breast cancer such as genetics, pollution, radiation or age. However, there may be other non-intuitive factors that contribute to breast cancer. The disclosed system and method will use social media post from users to collect data to suggest non-intuitive causes of disease.

Let's start with an example for demonstration only.

A researcher would like to select a set of possible cause of breast cancer and do formal experimental testing and verification. However, there are an infinite number of "possible causes." For example:
- Does driving a SUV lead to a higher chance of developing breast cancer?

- Does eating lunch at 1pm instead of 12pm lead to a higher chance of developing breast cancer?

Of course, researchers' time is limited and they will select the higher probable hypothesis to invest the time to do formal testing and verification. That means if the cause of breast cancer is non-intuitive, it will almost never be considered by researchers.

First, the disclosed system and method will perform an automatic patient search by performing the following steps:

Match email addresses from social media accounts with email addresses in discussion






If the system has a set of medical records (e.g. the researcher is an insurance company, or medical record company); the system can automatically match the social media accounts with the internal records (e.g. match by name, photo, email, location, phone, etc.). The system can use T(Pk,Hi) directly from the internal records, or a combination of all the above mentioned methods

Then, the disclosed system and method will perform a patient characteristic search Cj(Pk,Hi) for each patient "Pk" with or without health condition "Hi" before T(Pk,Hi) by searching for the following characteristics "Cj:"

Average time of the day for each meal;






Cj(Pk,Hi) can be in different formats:

Boolean (tongue rolling or not);

Enum (single, married, widow, etc.);

Integer/float (drinks per week, spouse's hair length).

Characteristic of spouse/partners (e.g. hair style, work, smoking, habit, car owned, etc.).


forums (e.g.; Identify a set of accounts for a particular individual;

Search for phrases from a predefined...